Service label in traefik_service_server_up metric is not human-readable

[ckoenig]

Welcome!

• Yes, I've searched similar issues on GitHub and didn't find any.
• Yes, I've searched similar issues on the Traefik community forum and didn't find any.

What did you do?

We are currently testing traefik v3 and as part of that we were investigating the metrics provided for Prometheus. We stumbled upon the following

# HELP traefik_service_server_up service server is up, described by gauge value of 0 or 1.
# TYPE traefik_service_server_up gauge
traefik_service_server_up{service="54033f984f82fa50",url="http://10.132.0.27:31012"} 1
traefik_service_server_up{service="5b667b985355cb4e",url="http://10.132.0.27:31331"} 1
traefik_service_server_up{service="6163188672cf989b",url="http://10.132.0.42:31447"} 1
traefik_service_server_up{service="616d0d8672d7da6f",url="http://10.132.0.42:31432"} 1
traefik_service_server_up{service="7a90018680d71149",url="http://10.132.0.42:31311"} 1
traefik_service_server_up{service="8caee5868b54566d",url="http://10.132.0.42:31157"} 1
traefik_service_server_up{service="8cbfdc868b62ba55",url="http://10.132.0.42:31100"} 1

We would have expected that the service label contains some human-readable value identifying the service affected. We also could not find any other metrics showing those IDs. So, there is also no way to join this with some other metric to retrieve a human-readable version.

The documentation mentions several metrics with a service label, e.g. traefik_service_requests_total. The service label for that metric is human-readable and not just some identifier.

What did you see instead?

Instead of seeing something like service="SERVICE@PROVIDER" we see service="SOMEID". This only seems to be the case for the traefik_service_server_up metric though.

What version of Traefik are you using?

$ kubectl describe deployment traefik  | grep Image
    Image:       docker.io/traefik:v3.0.0

What is your environment & configuration?

...
metrics:
  prometheus:
    service:
      enabled: true
    serviceMonitor:
      interval: 30s
...

Probably not relevant though.

If applicable, please paste the log output in DEBUG level

No response

[emilevauge]

Hey @ckoenig, thanks for submitting this issue. It might be a bug, we should see the serive name instead. If anyone from the community can have a look, that would be great :)

[KrishnaSindhur]

can I look into it!!

[stefanbankow]

Can confirm i'm experiencing similar behavior when running traefik through a docker-compose file. Reverting back to traefik:v2.6 fixes this issue and changes the traefik_service_server_up service label back to service-name@provider

[DrFaust92]

I can try and take a look